Authors: Eberhard Stickel, Otto Rauh
Tags: 1993, conceptual modeling
Absence of redundancy is an important quality of ER schemata. Adequate treatment of derivable data, in addition to observing normal forms, is necessary to achieve nonredundant databases. According to the source of their meaning to the user, there are two types of derivable data which should be treated differently. M-derivable data are redundant und ought to be removed from the ER schema, whereas s-derivable data add knowledge to the database and may be kept in the schema if they are distinguished from original data. However, there are some classes of derivable data, which are often brought into an ER schema unintentionally and are hard to detect. We introduce an important class of such data, compositions of relationship sets, and define them formally. In addition, we suggest methods which might help the database designer to find compositions in the schema.Read the full paper here: https://link.springer.com/chapter/10.1007/BFb0024358
